Guidance about the proper way to install the enclosure on the BlueBoat would be greatly appreciated, as the guide doesn’t seem to cover it (the one I found in the enclosure page).
That camera looks cool, but as you point out it’s not water proof. We tested this camera and were moving towards offering it on the Reef, but unfortunately found it was not possible to adjust the camera IP address or other configuration aspects such as controlling its position via Ethernet. Maybe that isn’t the case for the camera you’ve identified?
Putting your camera in a clear tube oriented vertically would allow you to look through the walls, however this is likely to introduce significant distortion. Additionally, the camera is likely to quickly overheat in this housing…
There is no specific guide to mounting cylindrical enclosures to the BlueBoat- this doesn’t seem like a common application? However an enclosure clamp and the numerous mounting points above and below water should make such a problem quite solvable…
The explorehd camera is the best fit for the BlueBoat, as the DWE OS extension let’s you lower the output bitrate to better suit the Wi-Fi bandwidth available to the system…
Regarding the camera you mentioned - I don’t quite understand the question - are you asking if we can connect to the A8 mini can connect to the BlueBoat OS? Also - the one you mentioned says “2-axis gimbal” which is probably not enough for our purpose.
The exploreHD has not gimbal configuration at all, so I’d have to build one myself, which is possible but time consuming.
However, your point about image distortion is important, so I was wondering whether you know about a different waterproof 3-axis gimbal camera that is suitable for this purpose.
What are you actually trying to do / what requirements do you have for your camera system? Are you trying to look at things above the boat, or in the water? Do you need to compensate for swell pitching and rolling the boat, or just point the camera in a general direction?
My goal is to have the camera looking underwater, and be stabilized to the point where the BlueBoat’s movement (induced by wind/wave/current) does not affect the camera’s orientation (other than translation, naturally).
Ok, if you want that to happen while the boat is driving then you likely do want some kind of waterproof enclosure around a camera with a gimbal, because otherwise the gimbal motors need to fight the drag of water flowing over the camera.
Ideally you’ll want an enclosure with a dome that the camera can sit in the focal point of - otherwise you’ll have to compensate for (or accept) distortions from the enclosure shape. One of our watertight enclosures with a dome on the submerged end could work well (and the cylindrical shape may reduce drag compared to a rectangular option, for example), but there are likely various possibilities for a suitable shallow depth enclosure, especially if you’re prepared to do some customisation/machining.
You could consider this camera, as it has h264 output and stabilizes. However, I don’t think it will stabilize while looking straight down, and the white balance may not be adjustable well for an underwater environment. The 3-inch enclosure required to use it would also add a lot of asymmetric drag on the Blue Boat which will impact efficiency.
If your goal is to observe the bottom from the Blue Boat, you may be much better suited to use a tow fish - this is already somewhat of the standard approach in the industry. I’ve been developing one that uses our upcoming 4k camera, but the existing low light USB camera with tilt mount would also work although at a lower quality. By using control surfaces to adjust drag to stabilize, roll and control depth, you are dramatically decoupled from the motion of the boat. I’ll have a guide detailing how to build and use one of these in the future, but we’re waiting for both that camera and the blue servo to be released before publishing. Happy to share more in the meantime if this seems like a better path for you. I’ve completed photogrammetry surveys of large areas of reef, hundreds of meters on a side, with 3.6 mm per pixel resolution.
What enclosure you’d select depends entirely on what camera you would put it in it! You’d probably want to pick the shortest length and diameter that can accommodate the camera. 3D printing a mount, and/or using the RAILS system to hold the internal pieces in place is the typical approach. You could use a clamp, or make your own with a 3D printer, to hold the enclosure to the keel holes on the BlueBoat, as mentioned in the integration guide previously linked.
The trickiest part to doing it yourself would be the cable - Ethernet is a bit easier to work with than USB, but both require twisted pairs of wires. You can’t pass the cable through the WetLink Penetrator bolts without removing the connector, so screw terminal to connector adapters are handy for making the connection both in the enclosure and in the BlueBoat hull.
As for your questions on the towifsh:
The Towfish is positively buoyant. This means that when the boat drives around after launch or in general, if the TowFish autopilot is disarmed, the hardware stays at the surface (and often upside down at this location.) This prevents it from fouling with underwater objects inherently!
A couple of fiberglass poles are a new addition - they act to keep the tow tether away from the BlueBoat props.
Once the TowFish is armed, if the BlueBoat is towing it forward, the control surfaces flip it over and bring it to depth. The 7m tether I’ve been using initially is sufficient length for the TowFish to achieve ~4m maximum depth. While not yet achieved, I’m working to get the TowFish to automatically go to a depth based on a static offset from the bottom, based on the Ping sonar measurements being taken by the BlueBoat. In this way, terrain following should be possible!
The camera in the housing is pointed straight down via a servo - this is the standard solution for the BlueROV2, and is available in both legacy and upcoming 4k Cam solution! Pitch compensation is turned on, so if the towfish pitches up and down the camera remains pointing straight down, acting as a 1 axis gimbal. Roll stability happens thanks to the control surfaces!
In my opinion, a TowFish is going to provide vastly better results than a hull mounted camera ever could, because the bottom is never going to be a consistent distance from your BlueBoat keel mounted camera. Additionally the motion of the BlueBoat can be so violent that it seems a gimbal would struggle to keep up…
You could indeed install a exploreHD camera on a TowFish - it does not have a servo to provide pitch compensation, so you could point it straight down but it may not be as stable as a camera inside the housing looking through the dome. The cable would simply route into the enclosure via the pre-installed WetLink Penetrator, connecting to the Raspberry Pi. The upcoming 4k Cam provides an endpoint for high quality jpegs, about 5-7 megabytes per image - this is much more data than the 1080p h264 video stream provides in a frame! My TowFish control extension, still in active development, embeds GPS and other data in the EXIF data of each captured image at 2Hz.
The TowFish is controlled by a dedicated Pi/Navigator - it is running ArduSub within BlueOS, and thinks the external servos are actually vertical thrusters!
You can review an example dataset here - many more assets are included, including logs from another boat running Omniscan 3D and RTK GPS, and drone / phone images and video. Please provide source reference if resharing - happy to answer any follow-up questions.
The BlueServo used by the TowFish doesn’t yet have a launch date, but the 4k Cam IP camera will be available in early September!
So, as I understand it - the current state of the TowFish is that full implementation of thruster based stabilization is still under development? Do you have a rough timeline when we can purchase the product as a whole?
Also, are you saying that the fully fledged TowFish is not planned to have pitch compensation?
And of course thank you for sharing the data. I can’t find any underwater videos from the TowFish, to get a feel of its stability. Can you point me to their location?
I’ve got no intention of making a TowFish variant that uses thrusters. Powering those thrusters with the voltage drop down the tether is likely not possible - adding a battery to the TowFish while keeping it positively buoyant would make things larger and heavier in air - something I’m trying to avoid. A majority of the components are already available for purchase - as mentioned a launch date for the servo has not yet been determined, and the 4k Cam will launch in early September. The electronics, tether, enclosure could all be purchased today! You can find my (very in development) CAD files here.
Pitch compensation comes from the tilt servo that holds the camera!
Do you have a rough cost estimate of the TowFish? And is there a bundled list of components and an assembly guide I can use as th go-to reference if we decide to choose this solution?
And considering that the TowFish is in relatively early stages, I am still lightly leaning towards going for the original Gimbal-based camera + Waterproof encolsure solution.
I have the BlueBoat with a mounting bracket installed on it. Is there an enclosure you would recommend and is there any guidance you can provide regarding the installation configuration/process? I want the camera to be full submerged, but sticking out as little as possible from the BlueBoat, and as little image distortion as possible.
The prices for some components aren’t finalized as they haven’t launched yet, but I would estimate a TowFish to cost just under $3k. I don’t have a final list or assembly guide yet, as the design is a ways from being finalized!
As mentioned, I don’t think you’ll find a gimbal based camera that can point straight down, and keep up with the motion of the BlueBoat. Without knowing what camera, it’s hard to pick the optimal enclosure, but a 3" one would likely fit most units. Using a dome for the enclosure makes sense, and the enclosure itself will likely protrude out of the water, with the dome just flush with the keel, so you can still set the boat on flat ground without the weight resting on the dome? Alternatively you may want to mount it so it can pivot down, deeper, when in use, and thus avoid potential bubbles near the surface.
I’d think you’ll need to design a mount that either clamps around the tube or attaches to an existing Blue Robotics tube clamp. It can attach to one hull via the subsurface keel mounting holes, or the cross bar?
As long as the camera communicates via ethernet or USB, and outputs H264 video, you should be all set to connect and handle the stream with BlueOS.
Is your goal photogrammetry? How rough will the water be when you’re surveying? The “KISS” solution here is the exporeHD camera mounted to the keel with its included mounting bracket… you may find that footage is usable depending on the sea conditions!
The goal is underwater animal detection and tracking, so I want the image frame to be as stable as possible (especially important for tracking).
Can you attack the links to the products you mentioned?
Again, the environmental conditions are going to determine your results more than anything, and the complication of a custom enclosure may not be necessary depending on this… And no matter what, in rough enough conditions, no camera, towed or gimbal, may be stable enough!
